Electro-thermal-mechanical Simulation and Reliability for Plug-in
Vehicle Converters and Inverters PI: Al Hefner (NIST)
May 14, 2013
Project ID # APE 026 This presentation does not contain any proprietary, confidential, or otherwise restricted information
• June 2011
• June 2014
• 40% Complete
Work Timeline
Budget
Barriers
• NIST- Electro-thermal modeling • UMD/CALCE – Reliability modeling • VTech – Soft switching module • Delphi – High current density module • Powerex – Module technology • NREL – Cooling technology
Partners
Overview
Need electro-thermal-mechanical modeling, characterization, and simulation of advanced technologies to: • Improve electrical efficiency • Improve package thermal performance and
increase reliability • Reduce converter cost
2
• Total project funding – $700K
• Funding received in FY11 – $ 200K
• Funding received in FY12 – $ 300K
• Funding expected in FY13 – $ 200K
3
Electrical
Mechanical
Thermal
• Electro-thermal interactions, • SOA and failure mechanisms, • Cooling system impacts.
• Inverter performance evaluation • Advanced topology design • Advanced device integration
Electro-Thermal
Electrical
Goal: Electro-Thermal-Mechanical Simulation
Reliability
• Reliable integration of advanced technologies
• System reliability evaluation.
• In-Vehicle applications: – Maintaining component health, – Predicting service needs, – Operation with partially degraded
capacity near component end-of-life.
Driving Cycles, Environmental Conditions Simulation Applications
3
Models, Parameter Determination
Electronic Component
Thermal Component
Mechanical Reliability
Relevance
4
Objective: Provide theoretical foundation, measurement methods, data, and simulation models necessary to optimize power module electrical, thermal, and reliability performance for Plug-in Vehicle inverters and converters.
FY 2013 Goals: 1) Analyze Viper SOA using dynamic electro-thermal simulation with
models including high voltage, high current parameter extraction 2) Develop Cross-Coupling TSP Measurement capability and use to validate
thermal coupling model within VTech Module Thermal Model 3) Develop Thermal Component Models for Air and Liquid Cooled Heatsinks
and include in electro-thermal simulation of Viper and VTech modules 4) Perform thermal cycle measurements to extract parameters for Physics-
of-Failure Models and use in Electro-Thermal-Mechanical Simulation 5) Develop electro-thermal models for advanced semiconductor devices
e.g., SiC MOSFETs and SiC JFETs and GaN diodes.
5
Milestones/Decision Points
Month/Yr Milestone
Aug. 12 (complete)
1) Used electro-thermal-mechanical simulations to validate measurement during fault conditions and evaluate thermal stresses in Viper module.
July 13 (Go,no-Go)
2) Incorporate Failure Models into Electro-Thermal Simulation using results of thermal cycling degradation and monitoring measurements on two DBC stacks.
Sept. 12 (complete)
3) Developed thermal-network-component models for representative cooling systems.
Oct. 12 (on hold)
4a) Used simulations to evaluate thermal stresses at module interfaces for VTech module, 4b) and use physics of failure models to calculate damage and evaluate impact on VTech module life.
Jan. 13 (on hold)
4c) Calculate increase in thermal resistance at interfaces in VTech module due to thermal cycling damage and use changing resistance in the thermal network during simulations.
Mar. 13 (complete)
5) Included liquid- and air-cooling thermal network component models in electro-thermal simulations of vehicle inverters.
June. 13 (ongoing)
6) Developed electro-thermal models for advanced semiconductor devices including SiC MOSFETs, SiC JFETs and GaN diodes.
Aug. 13 (ongoing)
7) Include advanced Wide-Bandgap semiconductor device models in simulations to optimize high current density, low thermal resistance, and soft-switching modules.
Mec
hani
cal
Ther
mal
El
ectr
ical
FY13 Tasks to Achieve Goals
2012
Oct
Nov
Dec
2013
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Develop Air and Liquid Cooling Heatsink Thermal Models
Go No/Go
Decision Point
Heatsink Thermal Models
Thermal Cycle measurements to extract Models for Physics-of-Failure Reliability
Viper Fault Model
Validate Thermal Cross Coupling in VTec Package Thermal Models
Model & Extract WBG Device Electro-Thermal Parameters
Electro-Thermal Models for Viper Fault Conditions
Viper Electro-thermal Simulation of SOA
Affix Package & Cooling Models to Electric Models
Failure Formula
Electro-Thermal Simulation of Device, Package, Heatsink
Viper SOA
6
Affix Viper Thermal Model to Fault Model
FY14 WBG Electro-Thermal Cost/Benefit
WBG Electric Models
VTec Thermal Model Incorporate Failure
Models into Electro-Thermal Simulation
Electro-Thermal-Mechanical Reliability Simulation
Affix Package and Cooling Models to WBG Models
7 7
Approach: Measurement, Modeling, and Simulation
Develop dynamic electro-thermal Saber models, parameter extractions, and validation of models for:
Silicon IGBTs and PiN Diodes Silicon MOSFETs and CoolMOSFETs SiC Junction Barrier Schottky (JBS) Diodes SiC MOSFETs
Develop thermal network component models and validate models using transient thermal imaging (TTI) and high speed temperature sensitive parameter (TSP) measurement:
Power Semiconductor Chip Package: Delphi VIPER and VTech Soft Switching modules Air and liquid cooling heatsinks
Develop thermal-mechanical degradation models and extract model parameters using accelerated stress and monitoring:
Stress types include thermal cycling, thermal shock, power cycling Degradation monitoring includes TTI, TSP, X-Ray, C-SAM, etc.
8
Application: Delphi Viper Module Double-Sided Cooling Model
A doubled-sided temperature-controlled heatsink that was developed for the Viper module. This heatsink uses a spring-loaded piston to apply a
controlled four kg compressional pressure to the device.
9 9
Method: Electro-Thermal Model for Double-Sided Cooling Viper Module
10 10
VG
VCE
.
.
.
IC
T
VGG
VCE
IC
T
RG
VClamp
Additional validation results given at FY12 PEEM Kickoff meeting.
IGBT
G
C
E
Tj
IGBT
N
P
TjPiN
Validation: Delphi-Viper Electro-thermal Semiconductor Models
11
Viper module 262 W steady state ANSYS simulation for double-cooled test fixture
Top Plate Temperature (0.05” from edge) 37 °C (ANSYS model) 40.4 °C (Measurement)
Piston Temperatures (0.05” from edge) 28 °C (ANSYS model) 31.7 °C (Measurement)
Device Temperature 69 °C (ANSYS model)
Viper module thermal test fixture
Viper Module
Top Cooling Plate (not shown) Piston
Water Cooling Fixture
Validation: Thermal Test Fixture Model
12
Validation: Thermal Network Component Model for Viper Module Package
Comparison of simulated and measured Junction Temperature (TSP), and Plate and Piston Temperatures (thermocouples) for a low power, long duration pulse.
Comparison of simulated and measured Junction Temperature (TSP) for short duration, high power pulses.
• Test fixture used to validate thermal model of Viper die, package, and interface to copper plates using TSP measurements.
• Test fixture modeled and compared with ANSYS and TSP.
200 W, 30 s Pulse
TSP Data ____ ____ Simulation
Thermocouples - - -Top Plate - - - Piston
2820 W
1175 W
1 ms Pulse ---- Simulations
TSP Data
12
13
Heat Sink Thermal Model
Heat Sink Thermal Model
Package Thermal Model
Q1(t)
Q2(t)
Qn(t)
Qtotal(t)
E(t)
y
V1 V2
P+
N-
N+
N++
Thermal Element
s
.
.
.
Package Thermal Model
junction
Package Thermal Model P+
Adiabatic Chip Heating (E•J)
I C [A
] Q
Tota
l [W
] V C
E [V
]
400 200
0
250
150
0 2 4 6 8 10
600
3k
1M
500k
0
T j [C
]
2k 1k 0
350
VClamp = 400 V T = 150 oC
Short Circuit Simulation
13
Method: Electro-Thermal Simulation Adiabatic Heating for Short Circuit Conditions
14 14
Required Tasks:
• Extend and Validate Viper IGBT model for high voltage, high current conditions • short pulse to reduce heating • characterize gate, collector and
common emitter R & L • vary Vg, Rg, Vc, and T to
characterize IGBT model • Thermal model validation using
longer short circuit pulse • Then, use model simulations to
analyze SOA performance
Demonstration: Viper Module Simulation for Short Circuit Conditions
IGB
T C
urre
nt [A
] IG
BT
Tj [C
]
Time [us]
IGB
T Vo
ltage
[V] 400
300
200
2.0k
1k
0
250
150 100
Vge = 9 V 15 V 11 V 13 V
1.0 2.0 3.0 4.0 6.0 7.0
Simulated
Measured
5.0
200
RG = 2.2 Ω Gate Pulse = 3 µs T = 150 oC
15
Heat Sink Thermal Model
Heat Sink Thermal Model
Package Thermal Model
Package Thermal Mode
junction
Package Thermal Model
Demonstration: Liquid-Cooled Heatsink Viper Module Thermal Simulation
PDissipated = 500 W, 540 ms
Fluid Temp
Chip Thermal Model
fluid properties, flow rate; number of fins; fin area; fin structure
Junction and chip nodes
DBC nodes
Sink nodes
PDissipated
heff = 2 W/(K·cm2)
Fluid Temp
chip area/location; thickness; density heat capacity; thermal conductivity
16
Sx1 Dx1
Sx2 Dx2
Dx3
Dx4
Module Components Q1
Q2
D1
D2 M2
Dx2
Dx4
Dx3 Sx2
Circuit Diagram
M1 Sx1 Dx1
Application: VTech Soft Switching Module
17
n1
n2
n3
Tj
Method: Thermal Network Component Models
17
18 18
. . .
VCE
I C
T
RG
VClamp
VG
VCE
IC
T
RG
VClamp
VG
Additional validation results given at FY12 PEEM Kickoff meeting.
Dra
in V
olta
ge [A
] 400
Time [us] 0
20
15
10
5
0 -5
300
200
100
0
-100 0.4 0.8 1.2 1.6 2
Current Voltage
…. Measured __ Simulated
Rg = 22 Ω
Dra
in C
urre
nt [A
] IG
BT
Cur
rent
[A]
IGB
T Vo
ltage
[A]
Time [us] 99 99.5 100 100.5 101
500 400 300 200 100
0 -100
500 400 300 200 100
0 -100
Rg = 4 Ω
Current Voltage
…. Measured __ Simulated
IGBT
G
C
E
Tj
IGBT
G
TjD
S
CoolMOS
N
P
TjPiN
Validation: VTech Module Electro-thermal Semiconductor Models
19
For the method to work, the IGBT has to dissipate a given power while the MOSFET remains off, and their gates
must be measured independently.
Method: Cross-Coupling TSP Measurement For VTech Module Paralleled IGBT/MOSFET
Qx1 IGBT
IGBT Gate E Sense
Shared IGBT Collector
& MOS Drain
IGBT Emitter MOS Source
MOS Gate S Sense
MOS1B
470 Ω
3 kΩ470 Ω
470 Ω
470 Ω 3 kΩ
60 V20 V
VceVgs
CH3
CH4
CH1
CH2 IGBT MOS
TSPIGBT = VCH1 – VCH2 TSPMOS = VCH3 – VCH4
The devices were chosen for having physical proximity, different power
dissipation ratings, and being thermally coupled through the same conductive layer on top of the DBC
20
Validation: Cross-Coupling TSP Measurement VTech Module Thermal Coupling Model
The IGBT was powered with a train of pulses at different duty cycles to
generate enough average heat to be sensed in the MOS vicinity.
This IGBT measurements were used to validate the thermal transient behavior
for the thermal stack model, and the MOS measurements were used to
validate the thermal coupling model between adjacent power devices.
Preliminary electro-thermal coupling model results for the MOS
measurements show a close correspondence in their behavior.
100 W @ 100% duty
80%
60%
40%
20%
10%
Simulated
Measured
21
y = 0.189x + 10.964
y = 0.1955x + 11.244
y = 0.1817x + 9.3553
8101214161820222426283032
0 10 20 30 40 50 60 70 80 90 100 110
Avg.
MO
S Te
mpe
ratu
re (C
)
Average Qx1 Power (W)
100 W
80 W
50 W
Peak Power
For each peak power test the duty cycle is increased from 0.1% to 99.9%, in 5% steps. NIST chilled water temperature determines initial point.
Analysis: Cross-Coupling TSP Measurements TSP Calibration for IGBT to MOSFET
22
Inverter Output
Pout : 20 kW fout: 60 Hz fsw : 20kHz Iout: 160 A Vbus: 114 V
Current detail Voltage detail
22
Analysis: Inverter Electro-thermal Simulation - VTech Module Electrical Waveforms
23 23
Analysis: Inverter Electro-thermal Simulation - VTech Module Electrical Waveforms
Current Crossing Current Crossing
Zero Crossing A variable timing scheme that uses a voltage sensing circuit to detect the
zero voltage crossing condition is used to determine the main switch turn-on time. The transformer current allows enough energy to discharge the main
device (Q1) voltage to zero prior to main device conduction, enabling the
zero-voltage switching condition.
24
Diode Voltage [V]
Dio
de C
urre
nt [A
]
0
50
100
150
200
250
0 0.5 1 1.5 2 2.5 3
25 oC
Si CoolMOS-
Diode
Si PiN Diode
Dio
de C
urre
nt [A
]
0 50
100
150
200 250
Diode Voltage [V] 0 0.5 1 1.5 2 2.5 3
125 oC
Si CoolMOS-
Diode
Si PiN Diode
Analysis of current sharing of paralleled Diodes (Si PiN, CoolMOS-body Diode, SiC JBS Diode)
Analysis of current sharing of paralleled Switches (Si IGBT and CoolMOS)
Analysis: Paralleled Si IGBT, CoolMOS, Diodes
25
Application: Package Reliability Prediction
Mean Temp.
Temp. Swing
Dwell Time
Tav,1 ΔT1 tdw,1
Tav,2 ΔT2 tdw,2
… … …
Tav,i ΔTi tdw,i
Physics-of-Failure Models • Coffin-Manson • Norris-Landzberg • Energy Partitioning • Strain-Range Partitioning
Technology Dependent Reliability
Models
Variable Ramp-Rate Thermal Cycling
High-speed Transient TSP Used to detect changes in thermal resistance of buried-interfaces caused by thermal cycling damage.
Degradation and Monitoring Design-of-Experiments
Reliability Simulations
Tav
ΔT
tdw
Nf (Tav, ΔT, tdw)
Multiple cycling parameters for each DBC stack construction.
26 26
• Validated Delphi Viper simulations for full range of short circuit fault conditions: collector voltages, gate-drive parameters, and initial temperatures.
• Electro-thermal-mechanical simulations used to evaluate thermal stresses
in Delphi Viper double-sided cooling power module for nominal and fault operating conditions.
• Performed a range of thermal cycling and thermal shock degradations to characterize mechanical reliability of two DBC stack types.
• Used new enhanced TSP measurement system to validate thermal cross-
coupling between die within VA Tech soft switching modules. • Performed full electro-thermal simulations and validations for VA Tech soft
switching module in propulsion inverter operation at Pout= 50 kW @ 20 kHz.
Summary
27 27
Future Work
• Include advanced Wide-Bandgap semiconductor device models in simulations to optimize high current density, low thermal resistance, and soft-switching modules.
• Develop electro-magnetic package/system interconnect models.
• Perform EMI simulations using electro-magnetic package/system interconnect models, electro-thermal semiconductor models and thermal-network-component models.
• Determine grid storage/inverter applications for bi-directional vehicle chargers and develop circuit simulation scripts for chargers operating in these conditions.
• Perform simulations and evaluate impact of advanced technology power semiconductors and module packages in bi-directional vehicle charger storage/inverter applications.
Top Related